Engineering Simulation: Is Your Analysis Fit For Purpose? Tim Morris – Chief Operating Officer, NAFEMS FEMCI Workshop 2

1. Engineering Simulation: Is Your Analysis Fit For Purpose? Tim Morris – Chief Operating Officer, NAFEMS FEMCI Workshop 2005

2. Outline • What does “fit for purpose” mean? • What needs to be fit for purpose? • How does NAFEMS fit in? • Current state of the practice • Ongoing activities • Future issues

3. Outline • What does “fit for purpose” mean? • What needs to be fit for purpose? • How does NAFEMS fit in? • Current state of the practice • Ongoing activities • Future issues

4. What does “fit for purpose” mean? • It depends who you are……….

5. What does “fit for purpose” mean? • For a Formula 1 engineer, speed is everything • “Even for a whole car aerodynamics model, we don’t need to perform any validation – we just know that it works. That’s good enough for us”

6. What does “fit for purpose” mean? • For a nuclear power safety engineer, reliability is everything • “We need to demonstrate overall reliability for the power station of 10-x. We can’t perform any tests. What is the reliability of an FEA calculation?”

7. Outline • What does “fit for purpose” mean? • What needs to be fit for purpose? • How does NAFEMS fit in? • Current state of the practice • Ongoing activities • Future issues

8. What needs to be fit for purpose? • Software (and hardware) • Analysts! • Procedures employed

9. Outline • What does “fit for purpose” mean? • What needs to be fit for purpose? • How does NAFEMS fit in? • Current state of the practice • Ongoing activities • Future issues

10. NAFEMS Background • Founded in 1983 “To promote the safe and reliable use of finite element and related technology” • Membership association • Not-for-profit organisation • International: 700 companies from around the world • Focused on engineering simulation technologies such as Finite Element Analysis and Computational Fluid Dynamics

11. NAFEMS Background • Board of directors formed from senior industrialists • Current chairman: Dr. Costas Stavrinidis, Head of Mechanical Engineering, ESTEC

12. NAFEMS Benchmark Studies “It has become possible for experienced designers, or novice engineers, with no knowledge of the finite element method (or desire to know) to model a structure and deliver answers. The Finite Element Method has become a black box, and no expert may be on hand to diagnose abuses of the system…… “Is NAFEMS Hitting the Right Target”, G. Davies, Imperial College, 1989

13. NAFEMS Benchmark Studies • …….NAFEMS has been trying therefore to ensure that codes have no mistakes; will produce respectable answers from respectable models; and are backed by a user community which can recognise faults and poor approximations when it sees them” • “Is NAFEMS Hitting the Right Target”, G. Davies, Imperial College, 1989

14. NAFEMS Benchmark Studies

15. Aims of NAFEMS • Primary purpose is to help members who are using engineering analysis to achieve better: • Collaboration with others in the industry • Innovation in the products that they develop • Productivity in their engineering design process • Quality of their simulations

16. Technical Working Groups • Education & Training Working Group • Computational Structural Mechanics Working Group • CFD Working Group • CAD/Integration Working Group • Analysis Management Working Group • Comprised of experts from industry and academia • Direct the technical activities of NAFEMS • Produce books, best practice guidelines etc.

17. Regional Steering Groups • Germany, Austria & Switzerland • UK • Italy • Nordic • Comprised of leading figures from industry, academia and software vendors • Direct the local activities of NAFEMS • Host seminars, meetings etc. • Provide feedback on the requirements of local NAFEMS members • North America • France • Spain and Portugal

18. Publications • Library of internationally acclaimed publications developed over the years including: • Primers • “How to…” Guides • “Why do...” Guides • Benchmarks • Issued to members as deliverables as they are developed

19. Events • Seminars in local regions • World Congress every two years • Highly focused events • Independent of vendors • Well supported by developers, industry and researchers

20. FENET Highlights • 110 participants - industry,academia, s/w • 12 European states • 4 years (Aug 2001- July 2005) • 2.2 M€ funding from EC • NAFEMS is the coordinator

21. FENET Rationale • Scale, depth & maturity of application of FE technology varies widely across industry • Benefits from sharing knowledge and experience • Current dissemination of “best practice” is not good

22. FENet -Technology Strategy Plan • Drivers in key industrial sectors • State of the art in relevant technical areas • State of practice in industry sectors • Research and technology development needs • Barriers to uptake of technology • Candidate topics for workshops/collaborative initiatives

23. Outline • What does “fit for purpose” mean? • What needs to be fit for purpose? • How does NAFEMS fit in? • Current state of the practice • Ongoing activities • Future issues

24. Are Most Analyses Fit For Purpose? • In recent years, a number of Round Robin exercises have been carried out. • Different analysts have submitted results to particular problems. • The results have been compared with each other, and with test. • The following slides show some example results.

25. Example Analysis I • Results from the Workshop on CFD in Ship Hydrodynamics, Gothenburg 2000 • Form factor prediction for the KRISO 300K tanker hull Form Factor = CT/CFO-1 • Variation Coefficient = 26.4% • Different results from the same code and turbulence model • Different results from different turbulence models • Variation increased at full scale Atkins, NAFEMS Seminar March 2004

26. Experimental Result Created by John Bergstrom, Luleå University of Technology Example Analysis II • Pressure recovery factor (efficiency) of a draft tube Qinetiq, NAFEMS Seminar March 2004

27. Example Analysis III EDF NAFEMS Seminar June 2003 MECA Project Concrete Cracking Nuclear Power Plant Prestressed Concrete Containment Vessel

28. Example Analysis III EDF NAFEMS Seminar June 2003 MECA Project Concrete Cracking Nuclear Power Plant Prestressed Concrete Containment Vessel

29. Joint Benchmark

30. Joint Benchmark – Sample Results

31. Joint Benchmark – Sample Results

32. Are Most Analyses Fit For Purpose? • We mustn’t jump to misleading conclusions. • Round Robin exercises rarely carried out using the quality control procedures that are usually adopted. • Nevertheless, the results do illustrate the need for adopting Best Practice Guidelines and working within a Quality Controlled set of procedures.

33. Outline • What does “fit for purpose” mean? • What needs to be fit for purpose? • How does NAFEMS fit in? • Current state of the practice • Ongoing activities • Future issues

34. What needs to be fit for purpose? • Software (and hardware) • Analysts! • Procedures employed

35. NAFEMS: Fit For Purpose Software • Continuing to develop Benchmarks in new areas

36. NAFEMS: Fit For Purpose Analysts • Registered Analyst Scheme

37. NAFEMS: Fit For Purpose Procedures • Quality Assurance Procedures for Engineering Analysis • Management of Finite Element Analysis – Guidelines to Best Practice • Quality System Supplement to ISO 9001 Relating to Engineering Analysis • SAFESA Guidelines • How to Undertake Contact and Friction Analysis • Workbook of Examples • ………..

38. Outline • What does “fit for purpose” mean? • What needs to be fit for purpose? • How does NAFEMS fit in? • Current state of the practice • Ongoing activities • Future issues

39. FENET Findings – Primary Issues • How can we determine and demonstrate the level of confidence that we have in our simulation results? • Integration of simulation into the overall design process • Requirement to more accurately represent real behaviour of engineering materials

40. FENET Findings – Confidence In Results • The key issue is all about validation: of the model, and of the results • How much confidence can you have that your results are “correct”? • Can you rely on simulation alone, without building physical prototypes? • If you perform tests to validate your simulation, how can you compare the results?

41. FENET Findings – Integration • The way in which simulation is used in the design process is rapidly changing. • Increasingly analysis is being used by “designers” as part of front loaded development: • “Toyota has slashed development costs and time by 30-40% and solves 80% of all problems before creating initial physical prototypes”1 • This brings up many issues concerning the requirements for training the wider pool of personnel who are to utilise simulation. • 1.“Enlightened Experimentation, The New Imperative for Innovation”, Stefan Thomke, Harvard Business Review, February 2001

42. FENET Findings – Materials Modelling Requirement for improved tools in many technical areas. E.g. • Representation of polymers • Turbulence modelling of fluids • Multiphysics • Fracture mechanics (for many materials including metals, composites, concrete etc) • Complex contact and friction in assemblies • Representation of welding • Current analysis capabilities often restricted by two factors: • Lack of suitable, robust, verified constitutive models • Lack of sufficient material data

43. FENET Findings – Aerospace Industry Sector • Annual Industry Meeting • (Plus Around 200 aerospace respondents to FENET FEA Survey) • Allowed ~50 Key Topics To Be Identified • Technology Readiness Levels, State of Practice, Priority Levels Established • Continuously Updated Throughout Project

44. FENET Findings – Aerospace Industry Sector • Most requirements derived from the business drivers: • Shorter development time and time-to-market. • Reduction in mass and power (fuel) consumption. • Increasing safety / responding to more stringent safety requirements. • Increasing quality and reducing production defects. • More integrated development processes, increasingly multi-disciplinary design and optimisation.

45. FENET Findings – Aerospace Industry Sector • Most important topics raised: • Shorter development time and time-to-market. • Need for knowledge based pre- and post-processors. • Too cumbersome interface between analysis and test. • Insufficient model validation and/or lack of test correlation leading to lack of confidence in results.

46. FENET Findings – Aerospace Industry Sector • Most important topics raised (continued): • Serviceability and reliability requirements to ensure that a product remains functional throughout its intended lifecycle, e.g. analysis that is required for circumstances which are not reproducible in physical testing: satellites in space environment, aircraft crashworthiness. Also derived from important business drivers such as avoiding warranty costs, cost of product recalls, large damage claims (in particular in US). • Consistent handling of uncertainty in analysis, i.e. modelling uncertainties, material property uncertainties, shape tolerances, realistic representative loads, in order to avoid worst-worst-case overdesign. This leads to need for established / accepted probabilistic approach(es). • The difficulties to obtain good material property data

47. FENET Findings – Aerospace Industry Sector • Tables available in Industry Reports • Information available for download from www.fe-net.org

48. Summary • What does “fit for purpose” mean? • What needs to be fit for purpose? • How does NAFEMS fit in? • Current state of the practice • Ongoing activities • Future issues